Welcome to Chuangxin Cloud!

Upload BOM My Account

Chuangxin Cloud

All Categories

Language

Home > The Blogs

Semiconductor April Song of Ice and Fire Top AI Performers Deliver Amid Industry Wide Price Surge

5/4/2026 11:17:18 PM

In April 2026, the semiconductor industry played out a remarkable "Song of Ice and Fire":

On the earnings front, Intel, TSMC, and SK Hynix all delivered historic results fueled by AI; on the pricing front, prices for memory, foundry services, and ICs rose across the board, with the list of price hikes continually expanding; in the application end market, AI computing power became more expensive and sought after, while consumer electronics struggled to survive through "downgrading" amid cost pressures, showing a stark divergence.

These three forces corroborate each other, collectively outlining the most authentic picture of the month: AI is no longer just a story; it is actively reshaping every link in the chain, from supply to demand.

Earnings Season: The "AI Story" Fully Materializes into "AI Performance"

In late April, leading U.S.-listed semiconductor companies successively disclosed their Q1 2026 earnings reports. Exceeding all expectations, it wasn't just AI chipmakers celebrating. Even traditional CPU giant Intel and analog chip behemoth Texas Instruments (TI) reported exceptionally strong results that were rare in their history.

1. Intel: Its Best Single-Day Performance in Nearly Four Decades

On April 23, Intel released its Q1 earnings: revenue of $13.6 billion, up 7% year-over-year, marking the sixth consecutive quarter of exceeding expectations; non-GAAP EPS was $0.29, while market expectations were a mere $0.01. What further shocked the market was its performance driver structure: Data Center and AI (DCAI) revenue reached $5.1 billion, a 22% year-over-year increase; AI-related businesses now account for 60% of the company's total revenue, growing 40% year-over-year.

Following the earnings release, Intel's stock surged about 20% in after-hours trading and closed up 23.6% the next day, marking its best single-day performance since October 1987-nearly four decades.

2. TSMC: Gross Margin Hits a Record High

As the "arms dealer" for AI chips, TSMC reported Q1 revenue of $35.9 billion, a 6.4% sequential increase; its gross margin soared to a record high of 66.2%. High-Performance Computing (HPC) accounted for 61% of revenue, with 3nm and 5nm capacity running at full utilization.

The company provided Q2 revenue guidance of $39.0-$40.2 billion and raised its full-year revenue growth forecast from "around 30%" to "above 30%". Capital expenditures are also expected to significantly exceed the previous two years, approaching the upper end of the $52-$56 billion range for the year. TSMC CEO C.C. Wei expressed confidence in AI's development trend, stating that semiconductor demand is expected to remain strong.

3. The "Explosive" Data from the Memory Duopoly

Samsung Electronics issued a preliminary earnings report on April 7: Q1 sales reached a record high of 133 trillion Korean Won, with operating profit soaring 7.5 times year-over-year to 57.2 trillion Won. The profit for a single quarter already surpassed the total for all of 2025. The memory business division contributed 95% of the total operating profit, with DRAM prices doubling and NAND prices rising nearly 70% in Q1.

Even more astonishing was SK Hynix. Its earnings report showed Q1 revenue of 52.58 trillion Won, a staggering 198% year-over-year surge; operating profit was 37.61 trillion Won, representing an operating margin of 72% and a net profit margin of 77%. This level of profitability even surpassed that of Nvidia and TSMC during the same period, making it the world's most profitable semiconductor company. HBM, high-capacity server DRAM, and enterprise SSDs were the real engines driving this miracle.

4. Texas Instruments: An Unexpected Surprise from the Analog Giant

Texas Instruments reported Q1 revenue of $4.83 billion, a 19% year-over-year increase, exceeding analyst consensus. Net profit reached $1.545 billion, up 31% year-over-year. The significant growth in these two key metrics signals a deepening recovery in the analog sector. The standout highlight was its Data Center business, which skyrocketed 90% year-over-year. The company's Q2 revenue guidance of $5.0-$5.4 billion also far surpassed analyst expectations. Weakness in PC and consumer electronics did not hinder TI's robust growth in the AI infrastructure sector.

5. STMicroelectronics: Revenue Up but Profits Down, AI Glow Dims

On April 24, STMicroelectronics released its Q1 earnings: revenue reached $3.10 billion, a substantial 23% year-over-year increase, though it declined 8.2% sequentially. Affected by one-time items including asset impairments, restructuring costs, and acquisition-related expenses, GAAP net profit was a mere $37 million, with an operating margin as low as 2.3%.

Although ST's automotive, industrial, and personal electronics businesses all rebounded, and management confirmed 2026 data center revenue will exceed $500 million, the precipitous decline in profitability and consecutive earnings misses have made it the leading company in this earnings season with the "weakest AI glow and the most fragile profitability."

After aligning the data, the main theme of the Q1 earnings reports becomes strikingly clear: All semiconductors related to AI (CPUs, advanced memory, high-performance analog/communication, foundry) have delivered their brightest historical report cards. However, chipmakers tied to consumer electronics and traditional automotive are consistently losing ground on profitability, facing the severe challenge of "increasing revenue without increasing profits." The landscape of fire and ice has never been so stark.

Price Hike Wave: From Memory to Foundry, the Entire Chain is "Buzzing with Price Increases"

If the earnings reports reflect the structural demand explosion brought by AI, then the comprehensive price hike wave that swept through in April reveals the deep tension in the global semiconductor supply side. This tension is not due to short-term capacity shortage but is the combined result of three forces: AI capacity absorption, soaring raw material costs, and geopolitical disruptions.

1. Major IDM Price Increase Announcements Take Effect

April was the effective date for price increase notices from numerous major IDMs (Integrated Device Manufacturers):

Texas Instruments (TI): Effective April 1, increases of 15%–85%, with industrial control products seeing the highest hikes.
Infineon: Effective April 1, attributed to the "massive demand" from AI data centers and rising raw material costs.
STMicroelectronics (ST): Raising prices across all product lines starting April 26, completing the price adjustment loop with Infineon in the same competitive arena.
NXP, onsemi, Diodes: All initiated price increases around April 1.

2. Foundry and OSAT: The Starting Point of Full-Chain Transmission

On April 1, Vanguard International Semiconductor (VIS) and Powerchip Technology (PSC) simultaneously raised quotes for mature process foundry services by approximately 15% and 15%–30%, respectively. United Microelectronics Corporation (UMC) plans to increase prices by about 10% in the second half of the year. Nexchip Semiconductor announced a unified 10% price increase effective June 1.

The direct reasons for foundry price hikes: TSMC and Samsung are continuously exiting 8-inch mature capacity, with global 8-inch capacity expected to shrink by about 10%. Simultaneously, rising gold and silver prices are driving up packaging costs, alongside increasing prices for specialty gases.

Foundry price increases have swiftly been passed down to fabless IC design companies. Driver IC design houses like ILITEK and Sitronix raised quotes by about 15%–20% in April, while power components (MOSFETs, IGBTs) saw a concurrent 10%–20% price increase.

3. Domestic Manufacturers: Over Ten Companies Follow Suit in Dense Succession

The actions of domestic Chinese semiconductor manufacturers were particularly dense during April's price hike wave. It was not merely "following the trend"; many companies had already initiated multiple rounds of price adjustments since the beginning of the year, forming the solid foundation of this wave.

The reasons for the price increases given by the companies are highly consistent: a substantial rise in core raw material costs like wafers and packaging materials, coupled with the capacity absorption effect brought by AI and the contraction of mature process supply. The industry is decisively shifting from the previous "price war" phase to a "profit recovery" stage.

In the MCU and power management fields, Puya Semiconductor acted earliest, raising prices for its general-purpose MCU products starting April 15. Fortior Tech and Jiejie Microelectronics had issued notices in mid-March, effective April 1. Novosense sent price adjustment letters to customers as early as March 23. On April 7, Nationalchip added the final piece, raising prices for its MCU and related chip products by 15% to 20%. Consequently, from April 1 to April 15, the price hike windows for over ten domestic MCU and power management design companies were almost seamlessly connected.

4. Memory: The Eye of the Storm, a Seller's Market

Entering the second quarter, memory price increases have continued to exceed expectations. DRAM is projected to rise an additional 58%-63%, and NAND flash by 70%-75%. The rise in NAND flash prices in Q2 has even surpassed that of DRAM, making it the new center of price hikes. Goldman Sachs has significantly raised its full-year 2026 DRAM/NAND flash price increase forecasts to 250%-280% and 200%-250%, respectively.

Accompanying the aggressive price hikes is the accelerated clearing of niche memory capacity. In April, Kioxia announced the complete phase-out of its 32nm/24nm/15nm 2D NAND and early BiCS 3D products, completely exiting the 2D NAND arena. Samsung also stopped accepting new orders for LPDDR4/LPDDR4X, shifting production lines towards HBM and advanced DDR5. The strategic withdrawal of these two giants will further widen the supply gap for niche products like small-capacity eMMC, SLC NAND, and LPDDR4, and amplify price increase expectations.

Furthermore, a subtle change occurred in the April market: spot prices for DDR4 in channels like Shenzhen's Huaqiangbei plummeted by 25% at one point, creating a rare divergence of "falling spot prices amidst rising contract prices." This precisely reflects how the AI-dominated supply structure is reshaping the market: weak demand for consumer-grade DDR4 and increased pressure on channels to destock, while prices for AI-related products like DDR5 and enterprise SSDs remain firm. Currently, the three major manufacturers still lock over 70% of their advanced capacity for HBM and DDR5, continuously squeezing consumer-grade capacity. Short-term spot price fluctuations do not alter the long-term tight-supply landscape. The short-term disturbance on the spot side does not change the seller's market nature; the strong pricing power in memory will persist.

Application End: The "Parallel Worlds" of AI and Non-AI

The shockwave of across-the-board component price hikes has finally torn a clear dividing line in the application end market: hardware and services related to AI are in a state of fervor, while consumer electronics and entry-level markets are showing clear signs of chill.

AI Track: The More Expensive, The Higher the Demand. Computing Inflation Fails to Dampen Demand.

Cloud service providers were the first to break the industry norm of "prices only go down, not up." On April 18, Alibaba Cloud and Baidu Intelligent Cloud simultaneously raised prices for their AI computing power and storage products, with increases of up to 34%. Tencent Cloud increased prices for some of its AI model services by over 4 times.

Price hikes have not discouraged customers. The leasing price for NVIDIA's H100 actually rose again in April, showing a rare market phenomenon where higher prices drive stronger demand. The order visibility for AI server manufacturers now extends into 2027, with long-term agreements signed with memory and CPU suppliers often amounting to billions of dollars. As SK Hynix frankly admitted in its earnings call, "it's not that we don't have capacity; it's that there truly isn't enough to go around."

Consumer Electronics: Under Cost Pressure, End Devices Start "Downgrading" to Survive.

In stark contrast to the AI fervor, the PC and smartphone markets are under the dual pressure of rising costs and weak demand. Lenovo raised prices for some of its mid-to-high-end business and gaming laptops by over 1,000 RMB; ASUS warned that price increases in some markets could reach 25%-30% in the next quarter.

Smartphone manufacturers are responding in their own ways. OPPO directly raised prices for its new model by about 300 RMB; Xiaomi adjusted retail prices for some of its products. More mid-range models are adopting a "downgrading" strategy-using previous-generation processors and reducing storage capacity-to maintain the critical 1999 RMB price point. As a salesperson for Honor helplessly stated, "If you need it, I suggest buying as soon as possible, as prices will go up again."

The most disadvantaged is the entry-level market. Omdia predicts the sub-$500 entry-level PC market will shrink by 35%, and global PC shipments in 2026 are expected to decline by approximately 12% year-over-year. These most price-sensitive consumers bear the brunt at the end of the raw material cost increase chain. Their choice is not to "buy the expensive option" but rather to "not buy at all."

Divergence is Truth. The application-end picture in April completely aligns with the divergence between AI and non-AI sectors seen in the earlier earnings reports. The semiconductor price hikes are not evenly distributed; instead, they are accelerating industry consolidation: computing inflation makes the strong stronger, while marginalizing the peripheral markets even faster.

Conclusion

In April 2026, the semiconductor industry told a clear story through multiple earnings reports, numerous price increase notices, and stark divergence in the end markets: AI has transformed from a story into performance, from expectation into reality, and from opportunity into a watershed.

The relentless climb in prices for memory, CPUs, and foundry services is not a simple cyclical recovery but rather the structural reshaping of the supply side by AI data centers. Those closest to AI enjoy the dividends of price increases and valuation premiums. Those stuck in the old battlefields of consumer electronics and the automotive industry endure the dual pressures of cost squeeze and profit decline.

Short-term fluctuations in the spot market and the shrinkage of the entry-level market have not shaken the hardest logic: as long as AI computing infrastructure construction does not slow down, the semiconductor seller's market will not end, and the trend of price increases and divergence will not reverse. This is not a summer of broad-based price increases, but rather a structural marathon. Those who can keep running will always be the frontrunners positioned on the AI supply side.